Personalizing exoskeleton-based upper limb rehabilitation using a statistical model: A pilot study

Abstract

Clinical studies have so far not been able to show if robotic therapy is superior to conventional methods. The personalization of robot-assisted therapy according to the individual motor deficits might contribute to reach this goal. Here we present a statistical approach to automatically personalize robotic rehabilitation. Our method uses different motor performance measures to estimate motor improvement and adapt the motor task within a treatment session. This approach was tested with a pilot sub-acute stroke patient and the outcome was compared to a similar patient who underwent conventional physical therapy. Pilot results showed better outcomes in clinical tests, kinematics and muscle activity for the subject who trained using the personalized robotic approach.